JP7535575B2 - Systems, methods, and computer program products for predicting, forecasting, and/or assessing tissue properties - Patents.com - Google Patents

Description

The present disclosure relates generally to diagnostic imaging and, in some non-limiting embodiments or aspects, to predicting, forecasting, and/or assessing later phases and/or tissue characteristics of medical images from selected phases.

Magnetic resonance imaging, or MRI for short, is an imaging method used especially in medical diagnostics to depict the structure and function of tissues and organs in the human or animal body.

In MRI, the magnetic moments of the protons in the examined subject are aligned in a basic magnetic field, so that there is a macroscopic magnetization along the longitudinal direction. This is then deflected from the rest position (excitation) by the incident radiation of a radio frequency (HF) pulse. The return from the excited state to the rest position (relaxation) or magnetization dynamics is then detected as a relaxation signal by one or more HF receiver coils.

In spatial encoding, rapidly switched gradient fields are superimposed on the base magnetic field. The captured relaxation signals, or the detected spatially resolved MRI data, are initially present as raw data in spatial frequency space and can be transformed into real space (image space) by a subsequent Fourier transformation.

In native MRI, tissue contrast is created by different relaxation times (T1 and T2) and proton densities.

T1 relaxation describes the transition of the longitudinal magnetization to its equilibrium state, T1 is the time required to reach 63.21% of the equilibrium magnetization before resonant excitation. It is also called the longitudinal relaxation time or spin-lattice relaxation time.

Analogously, T2 relaxation describes the transition of transverse magnetization to its equilibrium state.

MRI contrast agents exert their effect by modifying the relaxation times of structures that incorporate them. A distinction can be made between two groups of substances: paramagnetic and superparamagnetic. Both groups of substances have unpaired electrons that induce a magnetic field around individual atoms or molecules.

Superparamagnetic contrast agents result in a significant shortening of T2, whereas paramagnetic contrast agents result primarily in a shortening of T1. A shortening of T1 time results in an increase in signal intensity in T1-weighted MRI images, and a shortening of T2 time results in a decrease in signal intensity in T2-weighted MRI images.

The effect of the contrast agent is indirect, since it does not itself emit a signal, but instead only affects the signal intensity of the hydrogen protons in its vicinity.

Extracellular, intracellular, and intravascular contrast agents can be distinguished according to their pattern of spread within tissue.

An example of a superparamagnetic contrast agent is iron oxide nanoparticles (SPIO). Superparamagnetic contrast agents are commonly used as intravascular agents.

Examples of paramagnetic contrast agents are gadopentetate dimeglumine (trade names: Magnevist® and others), gadobenate meglumine (trade names: Multihance®), gadoteric acid (Dotarem®, Dotagita®, Cyclolux®), gadodiamide (Omniscan®), gadoteridol (ProHance®), and gadobutrol (Gadovist®), among other gadolinium chelates.

Extracellular, intracellular, and intravascular contrast agents can be distinguished according to their pattern of spread within tissue.

In CT, contrast agents absorb the x-ray radiation produced by the CT machine as it passes through the patient. In nuclear medicine, a form of molecular imaging, contrast agents contain radioactive atoms that decay to produce a signal that is detected by the imaging equipment. In hyperpolarized MRI, most nuclei are aligned with a magnetic field and injected into the patient for imaging.

Gadoxetic acid-based contrast agents are characterized by distinct uptake by hepatocytes, hepatic parenchymal cells, enrichment in functional tissues (parenchymal tissue), and enhanced contrast in healthy liver tissue. Cells of cysts, metastases, and most hepatocellular carcinomas no longer function like normal hepatocytes, do not take up the contrast agent or take it up very little, are not strongly depicted, and are consequently identifiable and localizable.

Examples of gadoxetic acid-based contrast agents are described in U.S. Pat. No. 6,039,931A, and they are commercially available, for example, under the trade names Primovist® or Eovist®.

The contrast enhancing effect of Primovist®/Eovist® is mediated by the stable gadolinium complex Gd-EOB-DTPA (gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid). DTPA forms a complex with the paramagnetic gadolinium ion that has extremely high thermodynamic stability. The ethoxybenzyl radical (EOB) is a mediator of hepatobiliary uptake of the contrast agent.

Primovist® can be used for the detection of tumors in the liver. The blood supply to healthy liver tissue is mainly achieved by the portal vein (vena portae), while the hepatic artery (arteria hepatica) supplies most primary tumors. After intravenous injection of a bolus of contrast agent, it is possible to observe a time delay between the signal onset of the healthy liver parenchyma and the signal onset of the tumor accordingly. Besides malignant tumors, benign lesions such as cysts, hemangiomas, and focal nodular hyperplasia (FNH) are frequently found in the liver. Proper planning of treatment requires that these be differentiated from malignant tumors. Primovist® can be used for the discrimination of benign and malignant focal liver lesions. By T1-weighted MRI, it provides information on the nature of the lesions. Differentiation is achieved by using the different blood supplies to the liver and tumor, as well as the temporal profiles of contrast enhancement.

The contrast enhancement achieved by Primovist® can be divided into at least two phases: the dynamic phase (including the so-called arterial, portal, and late phases) and the hepatobiliary phase, during which significant uptake of Primovist® into liver parenchymal cells has already occurred.

In the case of the contrast enhancement achieved by Primovist® during the distribution phase, what is observed is a typical perfusion pattern that provides information for the characterization of the lesion. By depicting the vascularization, it helps to characterize the lesion type and to determine the spatial relationship between the tumor and the blood vessels.

On T1-weighted MRI images, Primovist® produces clear signal enhancement in healthy liver parenchyma 10-20 minutes after injection (in the hepatobiliary phase), whereas lesions that contain no or only a few hepatocytes, e.g., metastases or moderately to poorly differentiated hepatocellular carcinoma (HCC), appear as darker areas.

Tracking the spread of contrast agent over time through the dynamic and hepatobiliary phases offers ample opportunity for detection and differential diagnosis of focal liver lesions, but the examination extends over a relatively long time span, during which motion by the patient must be avoided to minimize motion artifacts in the MRI images. Prolonged restriction of motion can be uncomfortable for the patient and difficult to achieve consistently in practice.

Non-limiting embodiments or aspects of the present disclosure are directed to improvements in diagnostic imaging tests, for example, in accelerating or shortening scan times (e.g., for detection and differential diagnosis of focal liver lesions using dynamic contrast-enhanced magnetic resonance imaging (MRI) and the like). Non-limiting embodiments or aspects of the present disclosure provide methods, systems, and computer program products for predicting, forecasting, and/or assessing later phases of medical images from selected phases (e.g., early phases, etc.), e.g., hepatic phases during hepatobiliary phases, thereby reducing data image acquisition times. Additionally or alternatively, parameters may be determined using data from imaging sequences that may replace and/or be used to synthesize later images, e.g., pharmacokinetic parameters of tissue volumes (e.g., voxels, etc.).

Although primarily described with respect to MRI of liver tissue using contrast agents, non-limiting embodiments or aspects of the present disclosure are not so limited, and methods, systems, and computer program products according to non-limiting embodiments or aspects may predict, forecast, and/or assess later phases of medical images from selected phases (e.g., early phases, etc.) of medical images of any organ acquired using any imaging modality with or without contrast agents.

According to some non-limiting embodiments or aspects, a method is provided that includes receiving a plurality of MRI images, the MRI images showing an examination area during a first time span; feeding the plurality of MRI images to a prediction model, the prediction model being trained by supervised learning to predict one or more MRI images showing the examination area during a second time span based on the MRI images showing the examination area during the first time span; generating one or more predicted MRI images showing the examination area during the second time span by the prediction; and displaying and/or outputting the one or more predicted MRI images and/or storing the one or more predicted MRI images in a data storage medium.

According to some non-limiting embodiments or aspects, a system is provided that includes a receiving device, a control and computing device, and an output device, the control and computing device being configured to prompt the receiving device to receive a plurality of MRI images, the received MRI images showing an examination area for a first time span, the control and computing device being configured to predict one or more MRI images based on the received MRI images, the one or more predicted MRI images showing the examination area for a second time span, and the control and computing device being configured to prompt the output device to display, output, or store the one or more predicted MRI images in a data storage medium.

According to some non-limiting embodiments or aspects, a computer program product is provided that includes a computer program that can be loaded into a memory of a computer, the computer program prompting the computer to perform the following steps: receiving a plurality of MRI images, the MRI images showing the examination area for a first time span; providing the received MRI images to a predictive model, the predictive model being trained by supervised learning to predict one or more MRI images showing the examination area for a second time span based on the MRI images showing the examination area for the first time span; receiving one or more predicted MRI images showing the examination area for the second time span as output from the predictive model; and displaying and/or outputting the one or more predicted MRI images and/or storing the one or more predicted MRI images in a data storage medium.

According to some non-limiting embodiments or aspects, there is provided a use of a contrast agent in an MRI method, the MRI method comprising the steps of: administering a contrast agent, the contrast agent spreading in an examination region; generating a plurality of MRI images of the examination region during a first time span; feeding the generated MRI images to a predictive model, the predictive model being trained by supervised learning to predict one or more MRI images showing the examination region during a second time span based on the MRI images showing the examination region during the first time span; receiving one or more predicted MRI images showing the examination region during the second time span as output from the predictive model; and displaying and/or outputting the one or more predicted MRI images and/or storing the one or more predicted MRI images in a data storage medium.

According to some non-limiting embodiments or aspects, a contrast agent for use in an MRI method is provided, the MRI method comprising the steps of: administering a contrast agent, the contrast agent spreading in an examination region; generating a plurality of MRI images of the examination region for a first time span; feeding the generated MRI images to a predictive model, the predictive model being trained by supervised learning to predict one or more MRI images showing the examination region for a second time span based on the MRI images showing the examination region for the first time span; receiving one or more predicted MRI images showing the examination region for the second time span as output from the predictive model; and displaying and/or outputting the one or more predicted MRI images and/or storing the one or more predicted MRI images in a data storage medium.

Further provided is a kit comprising an imaging agent and a computer program product according to an embodiment or aspect of the present disclosure.

Non-limiting embodiments or aspects of the present disclosure are more specifically elucidated below without distinguishing the subject matter of the embodiments (method, system, computer program product, use, imaging agent for use, kit). Conversely, the following elucidations are intended to apply analogously to all subject matter of all embodiments, regardless of the context (method, system, computer program product, use, imaging agent for use, kit) in which they occur.

When steps are described in a certain order in the description or claims, this does not necessarily mean that the embodiment or aspect is limited to the order described. Conversely, steps are also envisioned to be performed in different orders or otherwise in parallel with one another, unless one step is stacked on top of another step, which absolutely requires that the stacked steps be performed one after the other (however, this is clear in each individual case). Thus, the order described may be a preferred embodiment.

Non-limiting embodiments or aspects of the present disclosure shorten the time span of the examination of the subject in the generation of the MRI image. In some non-limiting embodiments or aspects, this is achieved by MRI images of the examination area of the subject being measured in a first time span (magnetic resonance measurement), and the measured MRI images are then used to predict, with the aid of a self-learning algorithm, one or more MRI images showing the examination area in a second time span. The actual magnetic resonance measurement of the subject is therefore limited to this disclosure and does not encompass the second time span. The MRI images showing the examination area during the first time span contain information that allows the prediction of the second time span.

The "test subject" may typically be a living organism, preferably a mammal, very specifically, preferably a human. The test area may be a part of the test subject, for example an organ or a part of an organ. In a non-limiting embodiment or aspect, the test area is the liver or a part of the liver of a mammal (e.g., a human).

The "examination region" (field of view, FOV), also called image volume, may in particular be the volume imaged in a magnetic resonance image. In general, it can be said that a three-dimensional field of view consists of one or more volume elements or voxels. If a two-dimensional FOV is considered, it can be said to consist of one or more voxels or one or more pixels (picture elements). The examination region can typically be defined by a radiologist (localizer), for example on an overview image. It is self-evident that the examination region can alternatively or additionally be defined automatically, for example based on a selected protocol.

An examination region is introduced into the base magnetic field. The examination region is subjected to an MRI procedure, which generates a number of MRI images indicative of the examination region during a first time span.

The term multiple means that at least two MRI images or measurements are generated, preferably at least three, and very specifically, preferably at least four MRI images or measurements.

The contrast agent that spreads to the examination region is administered to the subject. The contrast agent is preferably administered intravenously as a bolus in a weight-matched manner.

"Contrast agent" is understood to mean a substance or mixture of substances whose presence in a magnetic resonance measurement leads to an altered signal. Preferably, the contrast agent leads to a shortening of the T1 and/or T2 relaxation times.

Preferably, the contrast agent is a hepatobiliary contrast agent, such as Gd-EOB-DTPA or Gd-BOPTA.

In a particularly preferred embodiment, the contrast agent is a substance or mixture of substances having gadoxetic acid or a gadoxetate salt as the contrast enhancing active substance. Very particularly preferred is the disodium salt of gadoxetic acid (disodium Gd-EOB-DTPA).

Preferably, the first time span starts before or together with the administration of the contrast agent. It is advantageous when one or more MRI images are generated showing the examination area without contrast agent, since the radiologist can already obtain important information about the state of the subject's health in such images. For example, the radiologist can identify bleeding in such native MRI images.

The first time span preferably includes the contrast agent distributed in the examination region. Preferably, the first time span includes the arterial phase and/or the portal venous phase and/or the late phase in dynamic contrast-enhanced magnetic resonance imaging of the liver or part of the liver of the examination subject. The mentioned phases are defined and explained, for example, in the following publications: J.Magn.Reson.Imaging,2012,35(3):492-511,doi:10.1002/jmri.22833; Clujul Medical,2015,Vol.88 no.4:438-448,DOI:10.15386/cjmed-414; Journal of Hepatology,2019,Vol.71:534-542, http://dx.doi.org/10.1016/j.jhep.2019.05.005).

In a preferred embodiment, the first time span is selected such that an MRI image of the liver or a portion of the liver of the subject is generated,

Shows the examination area without contrast agent,

Shows the examination area during the arterial phase, when contrast agent spreads through the arteries into the examination area,

Indicates the area of examination during the portal venous phase, when contrast agent reaches the area of examination via the portal vein,

Shows the area examined during the late phase, when the concentration of contrast agent in arteries and veins decreases and the concentration of contrast agent in extravascular tissues and/or liver cells increases.

Preferably, the first time span begins within a time span of 1 minute to 1 second before the projection of the contrast agent or with administration of the contrast agent and continues for a time span of 2 minutes to 15 minutes, preferably 2 minutes to 13 minutes, even more preferably 3 minutes to 10 minutes from administration of the contrast agent. Since the contrast agent is very slowly excreted by the kidneys and/or biliary tract, the extended time span can extend up to 2 hours or more after administration of the contrast agent.

Because the contrast agent may spread at different rates in different examination subjects, the first time span may also be defined by the concentration of the contrast agent in different areas of the examination region. One possibility is depicted in FIG. 1.

Figure 1 shows a schematic representation of the time profile of the concentration of the contrast agent in the hepatic artery (A), hepatic vein (V) and healthy liver cells (P). The concentration is depicted in the form of signal intensity I in the stated areas (hepatic artery, hepatic vein, hepatic cells) in the magnetic resonance measurement as a function of time t. Upon intravenous bolus injection, the concentration of the contrast agent rises first in the hepatic artery (A) (dashed curve). The concentration passes through a maximum and then falls. The concentration in the hepatic vein (V) rises more slowly than in the hepatic artery and reaches its maximum later (dotted curve). The concentration of the contrast agent in healthy liver cells (P) rises slowly (continuous curve) and reaches its maximum only at a much later time point (not depicted in Figure 1). Several characteristic time points can be defined. At time point TP0, the contrast agent is administered intravenously as a bolus. At time point TP1, the concentration (signal intensity) of the contrast agent in the hepatic artery reaches its maximum. At time point TP2, the hepatic artery and hepatic vein signal intensity curves intersect. At time point TP3, the concentration (signal intensity) of the contrast agent in the hepatic vein passes through its maximum. At time point TP4, the hepatic artery and hepatocyte signal intensity curves intersect. At time point T5, the concentrations in the hepatic artery and hepatic vein have dropped to a level where they no longer cause measurable contrast enhancement.

In a preferred embodiment, the first time span includes at least time points TP0, TP1, TP2, TP3, and TP4.

In a preferred embodiment, MRI images are generated (by measurement) of at least all phases in the time spans TP0-TP1, TP1-TP2, TP2-TP3, and TP3-TP4.

It is assumed that in the time spans TP0-TP1, TP1-TP2, TP2-TP3, TP3-TP4, one or more MRI images are generated (by measurements) in each case. It is also assumed that a sequence of MRI images is generated (by measurements) during one or more time spans.

The term sequence refers to a time series, i.e. what is produced is multiple MRI images showing the examination area at successive points in time.

A time point is assigned to each MRI image or a time point can be assigned to each MRI image. Typically, this time point is the time point at which the MRI image is being generated (absolute time). Those skilled in the art will recognize that the generation of an MRI image uses a particular time span. What may be assigned to an MRI image is, for example, the time point of the start of acquisition or the time point of the completion of acquisition. However, it is also contemplated that any time point can be assigned to an MRI image (e.g., a relative time point).

Based on the time point, an MRI image can be positioned chronologically relative to another MRI image, and based on the time point of this MRI image, it is possible to establish whether a moment shown in an MRI image occurred chronologically before or after a moment shown in another MRI image.

Preferably, the MRI images are chronologically ordered in the series and plurality of MRI images, such that an MRI image showing an initial state of the examination region is placed before an MRI image showing a later state of the examination region in the series and plurality of MRI images.

The time span between two MRI images immediately following each other in the series and plurality of MRI images is preferably the same for all pairs of MRI images immediately following each other in the series and plurality of MRI images, i.e., the MRI images are preferably generated at a constant acquisition rate.

Based on the MRI images generated (by measurements) during the first time span, one MRI image is predicted or multiple MRI images showing the examination area are predicted during the second time span.

In some non-limiting embodiments or aspects, the second time span follows the first time span.

The second time span is preferably a time span within the hepatobiliary phase, preferably a time span beginning at least 10 minutes after administration of the contrast agent, preferably beginning at least 20 minutes after administration of the contrast agent.

The plurality of measured MRI images showing the examination region during the first time span are provided to a prediction model. The prediction model is a model configured to predict one or more MRI images showing the examination region showing a second time span based on the plurality of MRI images showing the examination region during the first time span.

In this context, the term "prediction" means that an MRI image showing the examination area during the second time span is calculated using an MRI image showing the examination area during the first time span.

The predictive model is preferably created with the aid of a self-learning algorithm in a supervised machine learning process. Learning is accomplished by using training data that includes a variety of MRI images of the first and second time spans.

During machine learning, a self-learning algorithm generates a statistical model based on training data. This means that examples are not simply memorized, but the algorithm "recognizes" patterns and regularities in the training data. Thus, the predictive model is able to assess unseen data. Validation data can be used to test the quality of the assessment of unknown data.

The predictive model is trained by supervised learning, i.e., multiple MRI images from a first time span are presented to the algorithm in succession and it is informed which of the MRI images in the second time span are associated with these multiple MRI images. The algorithm then learns the relationship between the multiple MRI images of the first time span and the MRI images of the second time span in order to predict one or more MRI images in the second time span for the unknown multiple MRI images of the first time span.

Self-learning systems trained by supervised learning are extensively described, for example, in C. Perez: Machine Learning Techniques: Supervised Learning and Classification, Amazon Digital Services LLC - Kdp Print Us, 2019, ISBN 1096996545, 9781096996545.

Preferably, the predictive model is an artificial neural network.

Such an artificial neural network comprises at least three layers of processing elements: a first layer with input neurons (nodes), an Nth layer with at least one output neuron (node), and an N-2th inner layer, where N is a natural number greater than 2.

The input neurons are responsible for receiving the digital MRI image as an input value. Typically, there is one input neuron for each pixel or voxel of the digital MRI image. There may be additional input neurons for additional input values (e.g., information about the examination region, information about the examination subject, and/or information about health conditions that are prioritized when generating the MRI image).

In such a network, the output neurons serve to predict one or more MRI images of a second time span given multiple MRI images of a first time span.

The processing elements of the layer between the input and output neurons are connected to each other in a predefined pattern with predefined connection weights.

Preferably, the artificial neural network is a so-called convolutional neural network (abbreviated as CNN).

A convolutional neural network can process input data in the form of a matrix. This makes it possible to use a digital MRI image, which is depicted as a matrix (e.g. width x height x color channels), as input data. In contrast, a regular neural network, for example in the form of a multi-layer perceptron (MLP), requires a vector as input, i.e., to use an MRI image as input, the pixels or voxels of the MRI image need to be rolled out successively in a long chain. As a result, a regular neural network cannot, for example, recognize an object in an MRI image independently of the object's location in the MRI image. The same object at different locations in the MRI image will have completely different input vectors.

CNNs essentially consist of alternating filters (convolutional layers) and aggregation layers (pooling layers), and finally one or more layers of "normal" fully connected neurons (dense layers/fully connected layers).

When analyzing sequences (sequences of MRI images), space and time can be treated as equivalent dimensions and processed, for example, by 3D folding. This is shown in the papers by Baccouché et al. (Sequential Deep Learning for Human Action Recognition; International Workshop on Human Behavior Understanding, Springer 2011, pages 29-39) and Ji et al. (3D Convolutional Neural Networks for Human Action Recognition, IEEE Transactions on Pattern Analysis and Machine Intelligence, 35(1), 221-231). Furthermore, it is possible to train different networks involved in time and space and finally merge the features, as explained in the publications by Karpathy et al. (Large-scale Video Classification with Convolutional Neural Networks; Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, 2014, pages 1725-1732) and Simonyan & Zisserman (Two-stream Convolutional Networks for Action Recognition in Videos; Advances in Neural Information Processing Systems, 2014, pages 568-576).

Recurrent Neural Networks (RNNs) are a family of so-called feed-forward neural networks that contain feedback connections between layers. RNNs allow for the modeling of sequential data by common use of parameter data through different parts of the neural network. The architecture for RNNs includes cycles that represent the influence of the current value of a variable on its own value at a future point in time, since at least a portion of the output data from the RNN is used as feedback to process subsequent inputs in the sequence.
Details can be gleaned, for example, from S. Khan et al.: A Guide to Convolutional Neural Networks for Computer Vision, Morgan & Claypool Publishers 2018, ISBN 1681730227, 9781681730226.

The training of a neural network can be carried out, for example, by the backpropagation method. In this context, what is required of the network is a mapping of a given input vector to a given output vector that is as reliable as possible. The mapping quality is described by an error function. The goal is to minimize the error function. In the backpropagation method, the artificial neural network is taught by modifying the connection weights.

In the trained state, the connection weights between the processing elements contain information about the relationship between the multiple MRI images of the first time span and the MRI images of the second time span, which can be used to predict one or more MRI images showing the examination area during the second time span for a new multiple MRI images showing the examination area during the first time span.

Cross-validation methods can be used to split the data into a training data set and a validation data set. The training data set is used for backpropagation training of the network weights. The validation data set is used to check the accuracy of predictions that the trained network can be applied to multiple unknown MRI images.

As already indicated, further information regarding the test subject, test area, and/or test conditions may also be used for training, validation, and prediction.

Examples of information about the subject are sex, age, weight, height, medical history, the effect and duration and amount of medications already taken, blood pressure, central venous pressure, respiratory rate, serum albumin, total bilirubin, blood glucose, iron content, lung capacity, etc. These can be collected, for example, from a database or electronic patient file.

Examples of information about areas of examination include pre-existing health conditions, surgeries, partial lung resections, liver transplants, liver iron, fatty liver, etc.

It is envisaged that the multiple MRI images showing the examination region during the first time span are subject to motion correction before they are fed into the predictive model. Such motion correction ensures that pixels or voxels of a first MRI image show the same examination region as corresponding pixels or voxels of a second MRI image downstream in time. Motion correction methods are described in EP 3118644, EP 3322997, US 20080317315, US 20170269182, US 20140062481, EP 2626718.

A system according to any non-limiting embodiment or aspect may perform a method according to any non-limiting embodiment or aspect.

The system includes a receiving device, a control and computing device, and an output device.

Although it is contemplated that the described devices are components of a single computer system, it is also contemplated that the described devices are components of multiple separate computer systems that are coupled together via a network for transmitting data and/or control signals from one device to another.

A "computer system" is a system for electronic data processing that processes data by programmable computational rules. Such a system typically comprises a "computer" and such a device comprises a processor for performing logical operations, and further peripheral devices.

In computer technology, a "peripheral" refers to any device that is coupled to a computer and that helps control the computer and/or acts as an input and output device. Examples are monitors (screens), printers, scanners, mice, keyboards, drives, cameras, microphones, loudspeakers, etc. Internal ports and expansion cards are also considered peripherals in computer technology.

Today's computer systems are often divided into desktop PCs, portable PCs, laptops, notebooks, netbooks, and tablet PCs, as well as so-called handhelds (e.g., smartphones), and all these systems may be utilized to carry out embodiments or aspects of the present disclosure.

Input into the computer system is accomplished by means of input means, such as, for example, a keyboard, mouse, microphone, and/or the like.

The system may be configured to receive a number of MRI images indicative of the examination area during a first time span, and generate (predict, calculate) one or more MRI images indicative of the examination area during a second time span based on these data and optionally further data.

The control and computation device serves to control the receiving devices, coordinate data and signal flow between the various devices, and compute the MRI images. It is envisioned that there may be multiple control and computation devices.

The receiving device serves for receiving a plurality of MRI images. The plurality of MRI images can be transmitted from a magnetic resonance system or read from a data storage medium, for example. The magnetic resonance system can be a component of the system. However, it is also envisaged that the system is a component of a magnetic resonance system.

The sequence of MRI images and optionally further data is transmitted from the receiving device to a control and computing device.

The control and computing device is configured to predict one or more MRI images based on a plurality of MRI images showing the examination area during a first time span, the predicted MRI images showing the examination area during a second time span. Preferably, loaded into the memory of the control and computing device is a prediction model used to calculate the MRI images of the second time span. The prediction model is preferably generated (trained) by supervised learning with the help of a self-learning algorithm.

Via the output device, the predicted MRI image can be displayed (e.g., on a screen), output (e.g., via a printer) or stored on a data storage medium.

Further non-limiting embodiments or aspects are set forth in the following numbered clauses:

Clause 1. A computer-implemented method comprising the steps of: acquiring measurement information associated with parameters of voxels of a patient's tissue measured at two or more time points, the two or more time points occurring before one or more properties of the tissue voxels are separable in an image generated based on parameters of the voxels measured at a single one of the two or more time points; and determining one or more properties of the tissue voxels based on the parameters of the voxels at the two or more time points.

Clause 2. The computer-implemented method of clause 1, wherein the one or more characteristics of the tissue voxels are further determined based on information associated with at least one of the patient and a health state of the patient.

Clause 3. The computer-implemented method of clauses 1 and 2, wherein one or more properties of a voxel of tissue are determined for at least one time point corresponding to two or more time points.

Clause 4. The computer-implemented method of any one of clauses 1-3, wherein one or more properties of a voxel of tissue are determined for two or more subsequent time points.

Clause 5. The computer-implemented method of any one of clauses 1 to 4, further comprising generating one or more images including one or more properties of the tissue voxels at the two or more subsequent time points based on the one or more properties.

Clause 6. The computer-implemented method of any one of clauses 1 to 5, further comprising: determining that the measurement information associated with a parameter of a voxel of the patient's tissue includes a threshold amount of measurement information associated with determining one or more characteristics of the voxel of the tissue; and controlling the imaging system to automatically stop acquiring the measurement information in response to determining that the measurement information includes the threshold amount of measurement information.

Clause 7. The computer-implemented method of any one of clauses 1 to 6, wherein the step of determining one or more properties includes providing measurement information associated with parameters of a voxel of the patient's tissue to a predictive model, the predictive model being trained by supervised learning to predict one or more properties of the voxel of the tissue based on the measurement information associated with the parameters at two or more time points.

Clause 8. The computer-implemented method of any one of clauses 1 to 7, wherein the step of determining one or more properties includes fitting a pharmacokinetic/pharmacodynamic (PK/PD) model of the tissue voxel to parameters of the tissue voxel measured at two or more time points, and determining one or more properties of the tissue voxel based on the PK/PD model fitted to the parameters of the tissue voxel measured at two or more time points.

Clause 9. The computer-implemented method of any one of clauses 1 to 8, wherein the step of determining one or more characteristics includes fitting a PK/PD curve of a plurality of pre-calculated PK/PD curves for a parameter to a parameter of a tissue voxel measured at two or more time points, and determining one or more characteristics of the tissue voxel based on the PK/PD curves fitted to the parameter at the two or more time points.

Clause 10. The computer-implemented method of any one of clauses 1 to 9, wherein the step of determining the one or more properties includes approximating a curve representing the one or more properties of the tissue voxel using a set of basis functions, fitting the approximated curve to parameters of the tissue voxel measured at two or more time points, and determining the one or more properties of the tissue voxel based on the approximated curve fitted to parameters of the tissue voxel measured at two or more time points.

Clause 11. The computer-implemented method of any one of clauses 1 to 10, wherein the step of determining one or more properties includes fitting a curve of a plurality of curves pre-calculated for the parameters using a set of basis functions to the parameters of the tissue voxel measured at the two or more time points, and determining one or more properties of the tissue voxel based on the curve fitted to the parameters of the tissue voxel measured at the two or more time points.

Clause 12. A system comprising one or more processors programmed and/or configured to: acquire measurement information associated with parameters of voxels of a patient's tissue measured at two or more time points, the two or more time points occurring before one or more properties of the tissue voxels are separable in an image generated based on parameters of the voxels measured at a single one of the two or more time points; and determine one or more properties of the tissue voxels based on the parameters of the voxels at the two or more time points.

Clause 13. The system of clause 12, wherein the one or more characteristics of the tissue voxels are further determined based on information associated with at least one of the patient and a health state of the patient.

Clause 14. The system of any one of clauses 12 and 13, wherein one or more properties of a voxel of tissue are determined for at least one time point corresponding to two or more time points.

Clause 15. The system of any one of clauses 12-14, wherein one or more properties of a voxel of tissue are determined for two or more subsequent time points.

Clause 16. The system of any one of clauses 12-15, wherein the one or more processors are further programmed and/or configured to generate one or more images including one or more properties of the tissue voxels at time points subsequent to the two or more time points based on the one or more properties.

Clause 17. The system of any one of clauses 12-16, wherein the one or more processors are further programmed and/or configured to determine that the measurement information associated with a parameter of a voxel of the patient's tissue includes a threshold amount of measurement information associated with determining one or more characteristics of the voxel of the tissue, and to control the imaging system to automatically stop acquiring the measurement information in response to determining that the measurement information includes the threshold amount of measurement information.

Clause 18. The system of any one of clauses 12-17, wherein the one or more processors are further programmed and/or configured to determine the one or more properties by providing measurement information associated with parameters of a voxel of the patient's tissue to a predictive model, the predictive model being trained by supervised learning to predict the one or more properties of the voxel of the tissue based on the measurement information associated with the parameters at two or more time points.

Clause 19. The system of any one of clauses 12-18, wherein the one or more processors are further programmed and/or configured to determine the one or more properties by fitting a pharmacokinetic/pharmacodynamic (PK/PD) model of the tissue voxel to parameters of the tissue voxel measured at the two or more time points, and determining the one or more properties of the tissue voxel based on the PK/PD model fitted to the parameters of the tissue voxel measured at the two or more time points.

Clause 20. The system of any one of clauses 12-19, wherein the one or more processors are further programmed and/or configured to determine the one or more characteristics by fitting a PK/PD curve of a plurality of pre-calculated PK/PD curves for the parameter to the parameter of the tissue voxel measured at two or more time points, and determining the one or more characteristics of the tissue voxel based on the PK/PD curves fitted to the parameter at the two or more time points.

Clause 21. The system of any one of clauses 12-20, wherein the one or more processors are further programmed and/or configured to determine the one or more characteristics by approximating a curve representing the one or more properties of the tissue voxels using a set of basis functions, fitting the approximated curve to parameters of the tissue voxels measured at two or more time points, and determining the one or more properties of the tissue voxels based on the approximated curve fitted to parameters of the tissue voxels measured at two or more time points.

Clause 22. The system of any one of clauses 12-21, wherein the one or more processors are further programmed and/or configured to determine the one or more characteristics by fitting a curve of a plurality of curves pre-calculated for the parameters using a set of basis functions to the parameters of the tissue voxel measured at the two or more time points, and determining the one or more characteristics of the tissue voxel based on the curve fitted to the parameters of the tissue voxel measured at the two or more time points.

Clause 23. A computer program product comprising at least one non-transitory computer-readable medium including program instructions that, when executed by at least one processor, cause the at least one processor to: acquire measurement information associated with parameters of voxels of a patient's tissue measured at two or more time points, the two or more time points occurring before one or more properties of the tissue voxels are separable in an image generated based on parameters of the voxels measured at a single one of the two or more time points; and determine one or more properties of the tissue voxels based on the parameters of the voxels at the two or more time points.

Clause 24. The computer program product of clause 23, wherein the one or more characteristics of the tissue voxels are further determined based on information associated with at least one of a patient and a health state of the patient.

Clause 25. The computer program product of any one of clauses 23 and 24, wherein one or more properties of a voxel of tissue are determined for at least one time point corresponding to two or more time points.

Clause 26. The computer program product of any one of clauses 23 to 25, wherein one or more properties of a voxel of tissue are determined for two or more subsequent time points.

Clause 27. The computer program product of any one of clauses 23 to 26, wherein the instructions further cause the at least one processor to generate, based on the one or more characteristics, one or more images including the one or more characteristics of the voxels of the tissue at the time points subsequent to the two or more time points.

Clause 28. The computer program product of any one of clauses 23 to 27, further comprising instructions to cause at least one processor to determine that the measurement information associated with a parameter of a voxel of the patient's tissue includes a threshold amount of measurement information associated with determining one or more characteristics of the voxel of the tissue, and to control the imaging system to automatically stop acquiring the measurement information in response to determining that the measurement information includes the threshold amount of measurement information.

Clause 29. The computer program product of any one of clauses 23-28, wherein the instructions cause at least one processor to determine one or more properties by providing measurement information associated with parameters of a voxel of the patient's tissue to a predictive model, the predictive model being trained by supervised learning to predict one or more properties of the voxel of the tissue based on the measurement information associated with the parameters at two or more time points.

Clause 30. The computer program product of any one of clauses 23-29, wherein the instructions cause at least one processor to determine one or more properties by fitting a pharmacokinetic/pharmacodynamic (PK/PD) model of the tissue voxel to parameters of the tissue voxel measured at two or more time points, and determining one or more properties of the tissue voxel based on the PK/PD model fitted to the parameters of the tissue voxel measured at two or more time points.

Clause 31. The computer program product of any one of clauses 23-30, wherein the instructions cause at least one processor to determine one or more characteristics by fitting a PK/PD curve of a plurality of pre-calculated PK/PD curves for a parameter to a parameter of a voxel of tissue measured at two or more time points, and determining one or more characteristics of the voxel of tissue based on the PK/PD curves fitted to the parameter at the two or more time points.

Clause 32. The computer program product of any one of clauses 23-31, wherein the instructions cause at least one processor to determine the one or more characteristics by approximating a curve representing one or more properties of a tissue voxel using a set of basis functions, fitting the approximated curve to parameters of the tissue voxel measured at two or more time points, and determining the one or more properties of the tissue voxel based on the approximated curve fitted to the parameters of the tissue voxel measured at the two or more time points.

Clause 33. The computer program product of any one of clauses 23-32, wherein the instructions cause at least one processor to determine the one or more characteristics by fitting a curve of a plurality of curves pre-calculated for the parameters using a set of basis functions to the parameters of the tissue voxel measured at the two or more time points, and determining the one or more characteristics of the tissue voxel based on the curve fitted to the parameters of the tissue voxel measured at the two or more time points.

Additional advantages and details of embodiments or aspects of the present disclosure are described in more detail below with reference to exemplary embodiments illustrated in the accompanying schematic drawings.

1 is a graph showing the time profiles of the concentration of contrast agent in the hepatic artery (A), hepatic vein (V), and hepatocytes (P), as well as the summed enhancement (S) for voxels containing each of the hepatic artery, vein, and hepatocyte tissues.

FIG. 1 illustrates a system according to a non-limiting embodiment or aspect.

1 is a flowchart of a non-limiting embodiment or aspect of a process for predicting, forecasting, and/or assessing tissue properties.

FIG. 1 shows MRI images of the liver during the dynamic and hepatobiliary phases.

FIG. 1 shows three MRI images (1), (2), and (3) showing the liver in a first time span, and an MRI image (4) showing the liver in a second time span.

FIG. 1 illustrates non-limiting embodiments or aspects of environments in which the systems, devices, products, apparatus, and/or methods described herein may be implemented.

1 is a flowchart of a non-limiting embodiment or aspect of a process for predicting, forecasting, and/or assessing tissue properties.

FIG. 1 shows a non-limiting example of a pharmacokinetic model of liver tissue.

FIG. 1 shows a composite MRI image (3) obtained by combining contrast information known from HBP for a liver-specific contrast agent with early phase MRI images (1) and (2) of the liver to generate a “plain white liver” image.

FIG. 1 shows non-enhanced T1w images (1), (2), (3) synthesized from measured contrast-enhanced images.

FIG. 13 shows images showing vascular enhancement separate from equilibrium uptake derived by subtracting one or more post-equilibrium updated images from one or more second post-injection images.

It is to be understood that the present disclosure may assume various alternative variations and step sequences, unless expressly noted to the contrary. It is also to be understood that the specific devices and processes illustrated in the accompanying drawings, and described in the following description, are merely exemplary and non-limiting embodiments or aspects. Hence, specific dimensions and other physical characteristics associated with the embodiments or aspects disclosed herein are not to be considered as limiting.

For purposes of the following description, the terms "end", "upper", "lower", "right", "left", "vertical", "horizontal", "top", "bottom", "transverse" and "vertical", as well as derivatives thereof, refer to the embodiment or aspect as they are oriented in the depicted figures. However, it should be understood that the embodiment or aspect may assume various alternative variations and step sequences, unless expressly noted to the contrary. It should also be understood that the specific devices and processes illustrated in the accompanying drawings and described in the following description are merely non-limiting example embodiments or aspects. Thus, specific dimensions and other physical characteristics associated with the embodiments or aspects disclosed herein should not be considered limiting, unless otherwise indicated.

As used herein, aspects, components, elements, structures, acts, steps, functions, instructions, and/or the like should not be construed as critical or essential, unless expressly described as such. Also, as used herein, the articles "a" and "an" are intended to include one or more items and may be used synonymously with "one or more" and "at least one". Furthermore, as used herein, the term "set" is intended to include one or more items (e.g., related items, unrelated items, combinations of related and unrelated items, etc.) and may be used synonymously with "one or more" and "at least one". When only one item is intended, the term "one" or similar language is used. Also, as used herein, the terms "has", "have", "having", or the like are intended to be open-ended terms. Additionally, the phrase "based on" is intended to mean "based at least in part on" unless expressly noted otherwise.

As used herein, the terms "communication" and "communicating" may refer to receiving, accepting, transmitting, forwarding, providing, and/or the like, information (e.g., data, signals, messages, instructions, commands, and/or the like). One apparatus (e.g., a device, a system, a component of a device or system, combinations thereof, and/or the like) in communication with another apparatus means that the one apparatus can receive information from and/or transmit information to the other apparatus, directly or indirectly. This may refer to a direct or indirect connection that is wired and/or wireless in nature. Additionally, two apparatuses may be in communication with each other even though the information transmitted may be altered, processed, relayed, and/or routed between the first and second apparatus. For example, a first apparatus may be in communication with a second apparatus even if the first apparatus passively receives information and does not actively transmit the information to the second apparatus. As another example, a first device may be in communication with a second device, even though at least one intermediate device (e.g., a third device located between the first device and the second device) processes information received from the first device and communicates the processed information to the second device. In some non-limiting embodiments or aspects, a message may refer to a network packet (e.g., a data packet and/or the like) that includes data. It should be understood that numerous other arrangements are possible.

As used herein, the term "computing device" may refer to one or more electronic devices configured to communicate directly or indirectly with or through one or more networks. A computing device may be a mobile or portable computing device, a desktop computer, a server, and/or the like. Furthermore, the term "computer" may refer to any computing device that includes the components necessary to receive, process, and output data, typically including a display, a processor, memory, input devices, and a network interface. A "computing system" may include one or more computing devices or computers. An "application" or "application program interface" (API) refers to computer code or other data stored on a computer-readable medium that may be executed by a processor to facilitate interaction between software components, such as a client-side front-end and/or a server-side back-end for receiving data from a client. An "interface" refers to a generated display, such as one or more graphical user interfaces (GUIs), with which a user may interact, directly or indirectly (e.g., through a keyboard, mouse, touch screen, etc.). Additionally, multiple computers, e.g., servers, or other computerized devices, such as an autonomous vehicle, including a vehicle computing system, communicating directly or indirectly within a network environment may constitute a "system" or "computing system."

It should be apparent that the systems and/or methods described herein may be implemented in the form of hardware, software, or a combination of hardware and software. The actual dedicated control hardware or software code used to implement these systems and/or methods is not a limitation of the implementation. Thus, the operation and behavior of the systems and/or methods are described herein without reference to specific software code, and it should be understood that software and hardware may be designed to implement the systems and/or methods based on the description herein.

Some non-limiting embodiments or aspects are described herein in connection with a threshold value. As used herein, satisfying a threshold value may refer to a value being greater than the threshold value, above the threshold value, higher than the threshold value, equal to or greater than the threshold value, less than the threshold value, less than the threshold value, lower than the threshold value, less than or equal to the threshold value, equal to the threshold value, etc.

Reference is now made to FIG. 1, which shows a schematic illustration of the time profiles of the concentration of contrast agent in the hepatic artery (A), hepatic vein (V), hepatocytes (P), and the sum (S) for a voxel containing all three types of tissue (e.g., each of the hepatic artery tissue, the hepatic vein tissue, and the hepatocyte tissue), and has already been described in detail herein above.

Referring also to FIG. 2, FIG. 2 illustrates a schematic diagram of a system according to a non-limiting embodiment or aspect. The system (10) includes a receiving device (11), a control and computing device (12), and an output device (13).

Referring also to FIG. 3, FIG. 3 is a flowchart of a method according to a non-limiting embodiment or aspect. The method (100) includes:

receiving a plurality of MRI images (110), the MRI images showing an examination region during a first time span;

A step (120) of providing a plurality of MRI images to a predictive model, the predictive model being trained by supervised learning to predict one or more MRI images showing the examination region during a second time span based on MRI images showing the examination region during a first time span;

generating (130) one or more predicted MRI images showing the examination area during the second time span by the predictive model;

and a step (140) of displaying and/or outputting the one or more predicted MRI images and/or storing the one or more predicted MRI images on a data storage medium.

Referring also to FIG. 4, FIG. 4 shows MRI images of the liver during the dynamic and hepatobiliary phases. In FIG. 4(a), FIG. 4(b), FIG. 4(c), FIG. 4(d), FIG. 4(e), and FIG. 4(f), the same cross-section of the liver at different time points is always depicted. The reference symbols entered in FIG. 4(a), FIG. 4(b), FIG. 4(d), and FIG. 4(f) apply to all of FIG. 4(a), FIG. 4(b), FIG. 4(c), FIG. 4(d), FIG. 4(e), and FIG. 4(f), and each of them is entered only once for the sole purpose of clarity.

Figure 4(a) shows a cross-section of the liver (L) before intravenous administration of hepatobiliary contrast agent. At time points between those depicted by Figures 4(a) and 4(b), hepatobiliary contrast agent was administered intravenously as a bolus. It reaches the liver via the hepatic artery (A) in Figure 4(b). The hepatic artery is therefore depicted with signal enhancement (arterial phase). The tumor (T), which is supplied with blood mainly via the artery, similarly stands out as a brighter (signal enhanced) area than the hepatic tissue. At the time points depicted in Figure 4(c), the contrast agent reaches the liver via the vein. In Figure 4(d), the venous vessels (V) stand out as a brighter (signal enhanced) area than the hepatic tissue (venous phase). At the same time, the signal intensity in healthy hepatic tissue, which is supplied with contrast agent mainly via the vein, continuously rises (Figures 4(c) → 4(d) → 4(e) → 4(f)). In the hepatobiliary phase depicted in FIG. 4(f), the hepatocytes (P) are depicted with signal enhancement, while the blood vessels and tumor no longer have contrast agent and therefore appear dark.

Referring also to FIG. 5, FIG. 5 exemplarily and diagrammatically illustrates how three MRI images (1), (2), and (3) showing the liver in a first time span are fed to a prediction model (PM). The prediction model calculates an MRI image (4) showing the liver in a second time span from the three MRI images (1), (2), and (3). The MRI images (1), (2), and (3) can represent, for example, the MRI images shown in FIG. 4(b), FIG. 4(c), and FIG. 4(d), and the MRI image (4) can represent, for example, the MRI image shown in FIG. 4(f).

Medical needs not met by existing imaging and/or image analysis systems include at least the following: identification and assessment of healthy tissue and/or different grades of pathological tissue, differentiation of different disease and/or pathological stages, improved imaging/scanning workflow, generation of cell-specific and/or cell-functional information, and functional information about hollow or tubular organs or organ systems.

Non-limiting embodiments or aspects of the present disclosure provide and/or improve at least the following: identification and assessment of healthy tissue and/or different grades of pathological tissue (e.g., widespread pathology such as tissue fibrosis/cirrhosis, inflammation, fatty infiltration, functional impairment/death, and/or the like, localized pathology such as benign or malignant tumors, and/or the like), differentiation of different disease and/or pathology stages (e.g., widespread pathology, localized pathology, and the like), imaging/scanning workflows (e.g., faster image generation and/or acquisition, and the like), generation of cell-specific and/or cell-functional information (e.g., cell function such as cell ability and rate to take up and secrete a particular drug or metabolite, and/or the like, amount of cell oxidation, expression of a particular cell antigen, channel/membrane proteins, and/or the like, molecular information, and/or the like), and/or functional information regarding hollow or tubular organs or organ systems (e.g., vascular system, bile duct system, vascular pressure, such as portal vein system, and the like). Non-limiting embodiments or aspects of the present disclosure provide voxel-specific estimates of one or more of arterial circulation volume, portal vein volume, venous volume, extracellular volume, bile duct volume, normal hepatocyte volume, adipocyte volume, fibrosis volume, Kupffer cell volume, stem cell volume, other hepatocyte volume, and/or metastatic or other lesion or non-hepatocyte volume.

6, which is a diagram of an example environment 600 in which the devices, systems, methods, and/or products described herein may be implemented. As shown in FIG. 6, environment 600 includes a contrast injector 602, an injector control and computing system 604, an injector user interface 606, an imager 608, an imager control and computing system 610, an imager user interface 612, an image analysis and computing system 614, an image analysis interface 616, an image and data review interface 618, a hospital information system 620, and/or cloud computing and off-site resources 622. The systems and/or devices of environment 600 may be interconnected by wired connections, wireless connections, or a combination of wired and wireless connections. For example, the systems and/or devices of environment 600 may be interconnected (e.g., communicate information and/or data, etc.) via a cellular network (e.g., a Long Term Evolution (LTE) network, a third generation (3G) network, a fourth generation (4G) network, a fifth generation network (5G), a code division multiple access (CDMA) network, etc.), a public land mobile network (PLMN), a local area network (LAN), a wide area network (WAN), a metropolitan area network (MAN), a telephone network (e.g., the public switched telephone network (PSTN)), a private network, an ad-hoc network, an intranet, the Internet, an optical fiber-based network, a cloud computing network, a short range wireless communication network, and/or the like, and/or a combination of these or other types of networks. The contrast injector 602, the injector control and computing system 604, and/or the injector user interface 606 may comprise a contrast injection system having respective software and/or hardware for setting one or more injection protocols and delivering one or more contrast agents to a patient according to the one or more injection protocols. In some non-limiting embodiments or aspects, the contrast injector 602, the injector control and computing system 604, and/or the injector user interface 606 may include a contrast injection system as described in U.S. Pat. Nos. 6,643,537 and 7,937,134 and published in International Application No. WO2019046299A1, the entire contents of each of which are incorporated herein by reference. In some non-limiting embodiments or aspects, the contrast injector 602, the injector control and computing system 604, and/or the injector user interface 606 may comprise a MEDRAD® Stellant FLEX CT Injection System, a MEDRAD® MRXperion MR Injection System, a MEDRAD® Mark 7 Arterion Injection System, a MEDRAD® Intego PET Injection System, a MEDRAD® Spectris Solaris EP MR Injection System, a MEDRAD® Stellant CT Injection System with a Certegra® Workstation, and/or the like.

The imager 608, imager control and computing system 610, and/or imager user interface 612 may comprise an imaging system having respective software and/or hardware for setting up imaging protocols and acquiring non-contrast and contrast-enhanced scans of a patient. In some non-limiting embodiments or aspects, the imager 608, imager control and computing system 610, and/or imager user interface 612 may include an MRI system (e.g., MRI based on T1, T2, TWI, PD, mapping (fat, iron), multi-parameter techniques, hyperpolarized MRI, MR fingerprinting, elastography, etc.), a computed tomography (CT) system, an ultrasound system, a single photon emission computed tomography (SPECT) system, a positron emission tomography-magnetic resonance (PET/MRI) system, a positron emission tomography-computed tomography (PET/CT) system, and/or other imaging diagnostic systems. In some non-limiting embodiments or aspects, the imaging device 608, the imaging device control and computing system 610, and/or the imaging device user interface 612 may include an imaging system such as that described in U.S. Patent Application No. 16/710,118, filed December 11, 2019, the entire contents of which are incorporated herein by reference. In some non-limiting embodiments or aspects, the imaging device 608, the imaging device control and computing system 610, and/or the imaging device user interface 612 may include a Siemens Healthineers Somatom Go CT system, a General Electric Company Signa MR system, and/or the like.

In some non-limiting embodiments or aspects, one or more of the contrast injector 602, the injector control and computing system 604, the injector user interface 606, the imaging device 608, the imaging device control and computing system 610, the imaging device user interface 612, the image analysis and computing system 614, the image analysis interface 616, the image and data review interface 618, the hospital information system 620, and/or the cloud computing and off-site resources 622 may include a computer system as described herein and/or one or more components and/or peripherals of a computer system.

A patient may include a living organism (e.g., a mammal, a human, etc.) that includes multiple tissues and/or organs with different types of cells (e.g., a liver containing hepatic parenchymal cells, Kupffer cells, immune cells, stem cells, etc.), afferent and/or efferent supply/circulatory systems (e.g., arterial, venous, lymphatic, or biliary systems, etc.), and/or different compartments/spaces (e.g., vascular spaces, interstitial spaces, intracellular spaces, etc.).

The contrast agent delivered to the patient by the contrast injection system may be selected or configured according to the type of imaging system used to scan the patient. The contrast agents may include gadolinium-based contrast agents (GBCA) (e.g., for use in MRI), iodine-based contrast agents (e.g., for use in CT), ultrasound contrast media (e.g., microbubbles, etc.), and/or other more uncommon contrast agents, such as iron-based agents (e.g., small or ultra-small superparamagnetic iron oxide, manganese-based, CO2, or other agents), blood pool agents (e.g., with long blood half-lives in the blood vessels), agents with predominantly renal or predominantly hepatobiliary excretion, agents with cellular uptake and organ- or cell-specific uptake, agents with organ- or cell-specific binding (e.g., no cellular uptake), agents with long-term "retention" (e.g., FDG), and/or the like. The contrast agents may be radioactive. The contrast agent may be cell marker specific, meaning that the contrast agent binds or interacts with a specific cell surface or internal marker. The contrast agent involved may be natural contrast, e.g., oxygen levels in hemoglobin. The contrast agent may be negative contrast, e.g., decreasing the hematocrit and thus red blood cell signal in blood or microbubbles, replacing blood with gas. In normal MRI, gas provides no signal, in CT, gas results in increased transmission and thus decreased Hounsfield units.

The contrast injection system may deliver a single agent (e.g., a single contrast agent delivered by itself) or multiple contrast agents combined at the same time (e.g., multiple parallel injections, injection of two fluids mixed together, or one after the other (e.g., multiple sequential injections)). The imaging system may perform a single image acquisition or scan at one or more time points (phases), multiple acquisitions at one or more time points (e.g., using MR mapping techniques, etc.), and/or acquisitions over a continuous imaging period.

In some non-limiting embodiments or aspects, the contrast injection may be delivered to one or more different locations or compartments, such as the venous vascular compartment, the arterial vascular compartment, the lymphatic vascular compartment, and/or the like.

Referring now to FIG. 7, FIG. 7 is a flow chart of a non-limiting embodiment or aspect of a process 700 for predicting, forecasting, and/or assessing tissue characteristics. In some non-limiting embodiments or aspects, one or more of the steps of process 700 may be performed (e.g., fully, partially, etc.) by image analysis and computing system 614 (e.g., one or more devices of image analysis and computing system 614). In some non-limiting embodiments or aspects, one or more of the steps of process 700 may be performed (e.g., fully, partially, etc.) by another device or group of devices separate from or including image analysis and computing system 614, such as imager control and computing system 610 (e.g., one or more devices of imager control and computing system 610), hospital information system 620 (e.g., one or more devices of hospital information system), cloud computing and off-site resource system 622 (e.g., one or more devices of cloud computing and off-site resource system 622), and/or the like.

As shown in FIG. 7, in step 702, process 700 includes delivering contrast to a patient. For example, a contrast injection system may deliver (e.g., inject) contrast to a patient. By way of example, contrast injector 602, injector control and computing system 604, and/or injector user interface 606 may set one or more injection protocols and deliver one or more contrast to a patient according to the one or more injection protocols.

In some non-limiting embodiments or aspects, the contrast agent may be delivered to the patient in a small, short, and relatively quick bolus to allow individual phases (e.g., arterial phase, portal venous phase, venous phase, etc.) to be differentiated. For example, the contrast agent concentration may be relatively homogenous for a short period of "steady state" before the contrast agent disperses in the body, allowing short-term steady-state imaging. In some non-limiting embodiments or aspects, the contrast agent may be delivered to the patient to delay or extend the injection time of the contrast agent, as described in U.S. Patent No. 9,436,989, the entire contents of which are incorporated herein by reference. In some non-limiting embodiments or aspects, the contrast agent may be delivered to the patient to achieve a desired MR contrast agent concentration in the blood and/or tissue, such that the MR fingerprint may include one or more contrast agent-related parameters, as described in U.S. Patent Application No. 16/462,410, filed November 21, 2017, the entire contents of which are incorporated herein by reference. In some non-limiting embodiments or aspects, the injection rate/delivery of the contrast agent may be performed as a single fast bolus (e.g., to distinguish between phases), as a single bolus to allow imaging in actual time from time X to time Y, as two separate image periods (e.g., 0.5-2 minutes and then 10 or 20 minutes, etc.), as a single slow bolus (e.g., with a visually indistinguishable normal phase, etc.), or as a dual bolus (e.g., in a sequence that includes delivery of a first bolus, waiting, imaging, delivery of a second bolus, and imaging, etc.).

7, in step 704, the process 700 includes acquiring measurement information associated with the patient's tissue. For example, the imaging system may set one or more imaging protocols and acquire non-contrast and/or contrast-enhanced scans of the patient's tissue to acquire measurement information associated with the patient's tissue (e.g., one or more images or scans of the patient's tissue, etc.) according to the one or more imaging protocols. By way of example, the imaging device 608, the imaging device control and computing system 610, and/or the imaging device user interface 612 (and/or the image analysis and computing system 614) may acquire measurement information associated with parameters of voxels of the patient's tissue measured at two or more time points (e.g., values of parameters of voxels of the tissue at two or more time points, images or scans of the tissue that include or indicate the parameters at two or more time points, etc.). In some non-limiting embodiments or aspects, the two or more time points occur before one or more characteristics of a voxel of the tissue are separable (e.g., separable from one or more other characteristics and/or noise in the measurement data and/or an image created from the measurement data) and/or distinguishable (e.g., distinguishable by the human eye) in an image generated based on parameters of the voxel at a single one of the two or more time points.

In some non-limiting embodiments or aspects, the measurement information may be acquired (e.g., scanned, imaged, etc.) at different (e.g., separate) times or consecutively and/or as continuous image acquisition across one or more phases before, during, and/or after contrast injection resulting in one or more of the following parameters/image acquisition phases: natural phase (e.g., pre-contrast), arterial phase, portal venous phase, venous phase, equilibrium phase, and in the case of hepatobiliary uptake and excretion, the hepatobiliary phase (HBP).

In some non-limiting embodiments or aspects, image acquisition or scanning to obtain measurement data may begin at a time point (e.g., an optimal time point, etc.) determined based on the patient's circulation time as described in European Patent Application No. 20161359.3, the entire contents of which are incorporated herein by reference.

In some non-limiting embodiments or aspects, measurement information including one or more parameters of a tissue voxel may be acquired using at least one of the following techniques: an MRI acquisition technique (e.g., T1, T2, TWI, PD, MRI based on mapping (fat, iron), multi-parameter techniques, hyperpolarized MRI, MR fingerprinting, elastography, etc.), a CT acquisition technique, an ultrasound technique, a SPECT technique, a PET/MRI technique, a PET/CT technique, another imaging technique, or any combination thereof.

In some non-limiting embodiments or aspects, the parameters of the tissue voxels measured by the imaging system may include at least one of the following: T1 weighted (T1w), T2 weighted (T2w), proton density weighted (PDw), diffusion weighted (DWI), x-ray absorption, T1 and/or T2 relaxation time shortening, change in x-ray absorption, tracer uptake/metabolism and radioregistration, one or more pharmacokinetic model parameters, or any combination thereof. For example, an image acquired by the imaging system may include or be indicative of the parameters of the tissue voxels at the time the parameters are measured by the imaging system.

7, in step 706, the process 700 includes determining one or more characteristics associated with the patient's tissue. For example, the image analysis and computation system 614 may determine the one or more characteristics based on parameters of the voxels at two or more time points. As an example, the image analysis and computation system 614 may generate one or more images that include or are indicative of one or more characteristics of the voxels of the patient's tissue based on measurement information including one or more parameters of the voxels at two or more time points (and/or images that include or are indicative of parameters of the voxels at two or more time points). In such an example, the image analysis and computation system 614 may synthesize, enhance, combine, and/or replace/subtract one or more images acquired and/or not acquired by an imaging system that acquired or measured measurement information including parameters of the voxels at two or more time points. For example, a single characteristic may be displayed as a grayscale image. As another example, two or more characteristics may be displayed as a color overlay on a grayscale image.

In some non-limiting embodiments or aspects, one or more characteristics (and/or one or more images including or indicative of one or more characteristics) may be determined for a time point and/or time period corresponding to (e.g., at the same time as) one or more of the two or more time points. In some non-limiting embodiments or aspects, one or more characteristics (and/or one or more images including or indicative of one or more characteristics) may be determined for a time point and/or time period subsequent to (e.g., following) the two or more time points.

In some non-limiting embodiments or aspects, the characteristics associated with the patient's tissue may include at least one of the following: concentration of contrast agent in the artery of the voxel (e.g., in the hepatic artery (A)), concentration of contrast agent in the vein of the voxel (e.g., in the hepatic vein (V)), concentration of contrast agent in the cell of the voxel (e.g., in the hepatic cell (P)), aggregate concentration of contrast agent in the artery, vein, and cell of the voxel (e.g., aggregate concentration of contrast agent in the hepatic artery (A), hepatic vein (V), and hepatic cell (P) of a liver tissue voxel), one or more of a pharmacokinetic parameter associated with the movement of contrast agent through the tissue space of the voxel, or any combination thereof. In some non-limiting embodiments or aspects, the characteristics associated with the patient's tissue may include at least one characteristic not associated with the injected contrast agent, including, for example, electron density, hydrogen density, T1, T2, apparent diffusion coefficient (ADC), or any combination thereof.

In the early days of medical imaging, data acquisition and image creation were performed by x-rays passing through the patient and being absorbed by a luminescent screen coupled with a photographic film that was developed to create an image. The film was then read by a radiologist to perform the diagnosis. With current electronic and computer technology, there are multiple imaging modalities with different ways of acquiring 2D or 3D arrays of data (over time (4D)) and converting this data into an image, meaning a human understandable 2D, 3D, or 4D data representation. A fifth dimension can also be mentioned, which is the ability to collect multiple parameters of data at each time point for each voxel in the patient region of interest. As further shown in FIG. 7, exemplary but non-limiting data analysis steps include data analysis 711, which turns acquired data or images into a human readable image, data analysis 712, which uses, for example, two or more images to create an image that did not exist before in a human readable form. , for example, for review by an involved medical professional, or may include data analysis 713 that may inform or recommend to the reader or make a diagnosis. Computer-aided diagnosis algorithms are examples of this data analysis. For example, the image analysis and computation system 614 may perform at least one of the following techniques: window & level techniques, gamma correction techniques, filtered backprojection techniques, iterative reconstruction techniques, model fitting techniques, dictionary modes of curve fitting techniques (e.g., MR fingerprinting, etc.), spatial, temporal filtering and enhancement techniques, feature extraction techniques using CAD or AI, fractal analysis techniques, motion correction techniques, flexible registration techniques, parameter quantification techniques, 2D quantification of single images, and the like. One or more characteristics (and/or one or more images including or indicative of the one or more characteristics, and/or diagnostic information associated therewith) may be determined based on measurement information including parameters of voxels at two or more time points using AI or machine learning based analysis, 2D AI or machine learning based analysis techniques of time sequences of voxels, 3D AI or machine learning based analysis techniques of time sequences of images, AI or machine learning based noise reduction and sharpening techniques, AI or machine learning based contrast amplification techniques, or any combination thereof.

In some non-limiting embodiments or aspects, the image analysis and computation system 614 (and/or the imager control and computation system 610) may determine that the measurement information associated with parameters of the patient's tissue voxels includes a threshold amount of measurement information associated with determining one or more characteristics of the tissue voxels (e.g., a sufficient amount of information and/or data to determine one or more characteristics and/or one or more images including the one or more characteristics) and control the imaging system (e.g., imager 608, etc.) to automatically stop acquiring the measurement information (e.g., stop scanning or imaging the patient, etc.) in response to determining that the measurement information includes the threshold amount of measurement information. In some non-limiting embodiments or aspects, the threshold amount of measurement information associated with determining one or more characteristics of the tissue voxels may optionally include a threshold amount of information associated with providing one or more selected or predetermined diagnoses for the patient, with a predefined level of parameter derivative, prediction confidence, margin of error.

In some non-limiting embodiments or aspects, the image analysis and computation system 614 can generate or create one or more images of voxels of the patient's tissue based on the measurement information including parameters of the voxels at two or more time points (and/or one or more images including or indicative of parameters of the voxels at two or more time points) (e.g., one or more composite or enhanced images, one or more non-enhanced/contrast-enhanced/tracer images, one or more T1w/T2w/PDw images (and/or variants thereof, e.g., with/without fat-sat, with inversion pulse, etc.), one or more images of arterial/portal/venous/equilibrium/HB phase, one or more images of MRI sequential acquisition, one or more non-enhanced/contrast-enhanced/tracer x-ray absorption images, one or more arterial/portal/venous/equilibrium/or sequential acquisition x-ray absorption images, etc.). As an example, the image analysis and calculation system 614 may synthesize one or more images of the HBP from measurement information including parameters of voxels at two or more time points measured in one or more phases occurring prior to the HBP. As an example, the image analysis and computation system 614 may combine one or more images by comparing (e.g., adding, subtracting, multiplying, etc.) contrast agent information known from HBP for a liver-specific contrast agent with one or more images of an early phase of the liver (e.g., one or more images measured or acquired in the arterial, portal, and/or venous phases, etc.) to generate a “plain white liver” as shown in FIG. 9 and shown in detail in European Patent Application No. 19197986.3, filed September 18, 2019, and International Patent Application No. PCT/EP2020/075288, filed September 10, 2020, the entire contents of each of which are incorporated herein by reference. By way of example, the image analysis and computation system 614 may synthesize a non-enhanced T1w image from one or more measured contrast-enhanced images (e.g., in the liver, etc.) as shown in FIG. 10 and described in further detail in European Patent Application No. 19201919.8, filed October 8, 2019, the entire contents of which are incorporated herein by reference. By way of example, the image analysis and computation system 614 may synthesize a non-enhanced T1w image from one or more measured contrast-enhanced images (e.g., in the liver, etc.) as shown in FIG. 10 and described in further detail in European Patent Application No. 19201937.0, filed October 8, 209, and U.S. Provisional Patent Application No. 62/943,980, filed December 5, 2019, to enable high-contrast, high-precision imaging of tissues and tumors with reduced amounts of contrast agent. The image analysis and calculation system 614 may synthesize an equilibrium phase contrast-enhanced T1w image (e.g., a virtual blood pool contrast image, etc.) as described in European Patent Application No. 20167879.4, filed April 3, 2020, the entire contents of which are incorporated herein by reference. As an example, the image analysis and calculation system 614 may synthesize an image from a conventional gadoxetic acid-enhanced image without gadoxetic acid-induced contrast changes in liver parenchymal cells during the transition or late phase. As an example, the image analysis and calculation system 614 may synthesize a "pseudo image" based on a set of conventionally acquired mapping sequences that includes information about the complete relaxation curve of the tissue over the pharmacokinetic time frame during which the substance (e.g., contrast agent) passes through the tissue.

By way of example, a dual injection protocol of contrast agent may be used as shown in FIG. 11 and described in further detail in European Patent Application No. 19201919.8, filed October 8, 2019, European Patent Application No. 19202711.8, filed October 11, 2019, and European Patent Application No. 20173089.2, filed May 6, 2020, the entire contents of each of which are incorporated herein by reference. The first injection of the dual injection may be performed while the patient is not in the scanner. The first injection may be, for example, gadoxetic acid if the study is to be an MRI study. The patient remains outside the scanner while the contrast agent equilibrates in the body and is taken up by the target tissue. The patient is placed in the scanner and one or more images may be taken of the equilibrium uptake, as well as images that are not affected by the uptake. A second contrast injection of the dual injection is then given and one or more additional images are taken. These second post-injection images may be timed to allow visualization and/or measurement of vascular enhancement, for example, in addition to the equilibrium uptake imaging. As an example, the image analysis and computation system 614 may subtract one or more post-equilibrium uptake images from one or more second post-injection images to derive an image showing vascular enhancement separate from the equilibrium uptake at the corresponding time point. The second contrast injection may be the same type of contrast agent as the first contrast injection, or it may be a different type of contrast agent, e.g., gadobutrol, which has different, and in some ways advantageous, properties relative to gadoxetic acid, since uptake has already occurred and been imaged. In some non-limiting embodiments or aspects, two different contrast agents may be mixed or injected simultaneously. For example, a contrast agent with properties that make it preferred as a vascular contrast agent may be mixed with a contrast agent that has cell or tissue specific marker or contrast agent properties that are not as useful as a vascular phase contrast agent.

7, in step 708, the process 700 includes analyzing one or more characteristics associated with the patient's tissue. In some non-limiting embodiments or aspects, a radiologist and/or other trained professional may read one or more images including one or more characteristics associated with the patient's tissue to derive diagnostic information from the one or more images. In some non-limiting embodiments or aspects, the image analysis and computing system 614 may analyze one or more characteristics associated with voxels of the patient's tissue, and/or one or more images, one or more characteristics associated with the patient's tissue to derive diagnostic information from the one or more images. For example, the image analysis and computation system 614 may provide the other characteristic (and/or one or more images containing or exhibiting the one or more characteristics) to a classification model that has been trained by supervised learning to classify tissue associated with one or more diagnoses based on the other characteristic (and/or one or more images containing or exhibiting the one or more characteristics). By way of example, the image analysis and computation system 614 may determine one or more diagnoses for the tissue using one or more of the techniques described herein above with respect to step 706 used to determine one or more characteristics of the tissue.

In some non-limiting embodiments or aspects, the diagnostic information may include an identification of a state of a voxel of the patient's tissue (e.g., healthy tissue, unhealthy tissue, a type of unhealthy tissue, a type of disease associated with the tissue, etc.), one or more reasons for the identification of the state of the voxel of the patient's tissue, or any combination thereof.

As shown in FIG. 7, in step 710, process 700 includes providing a diagnosis associated with the patient's tissue. For example, a radiologist and/or other trained professional may provide diagnostic information associated with the patient's tissue. As an example, image analysis and computing system 614 may provide diagnostic information associated with the patient's tissue.

7, in some non-limiting embodiments or aspects, at step 720, additional information regarding the patient and/or the patient's health condition may be input and used in one or more of stages 702-710 to influence or influence the selection between options at each stage and/or the information and/or data determined and/or generated at each stage. For example, the additional information regarding the patient and/or the patient's health condition may be input into the injection protocol, imaging protocol, measurement information, one or more characteristics (and/or one or more images including or indicative of one or more characteristics), diagnostic information, and/or processes, algorithms, machine learning models or neural networks, model fitting analyses, and/or the like used to determine a diagnosis. By way of example, the additional information regarding the patient and/or the patient's health status and/or desired conditions for the test may include at least one of the following: the patient's height, the patient's weight, the patient's age, the patient's sex, the patient's heart rate, the patient's cardiac output, the patient's clinical condition, the patient's bilirubin level, the desired rate of change , rise, and/or plateau level of the concentration of the contrast agent delivered to the patient, or any combination thereof. Additional examples of patient-specific data can be found in U.S. Patent Nos. 5,840,026 and 9,616,166, the entire contents of each of which are incorporated herein by reference.

The data analysis steps 711, 712, 713 and associated computer hardware and/or software, e.g., image analysis and computing system 614, may also analyze and use additional data 720 about the patient, which may be obtained, for example, from a hospital information system 620 and/or other cloud computing data stores 622.

Referring now to FIG. 8, FIG. 8 shows a simplified schematic diagram of voxels 800 (e.g., volume elements, etc.) of a patient's organ being imaged. The schematic diagram represents the liver, but generally applies to other organs with appropriate adaptations or simplifications. The liver is a relatively complex organ, as is well known in medicine and physiology.

Each of the example elements may occupy some fractional volume of the voxel. The liver receives a dual blood supply from the hepatic portal vein 304 and the hepatic artery 302. The hepatic portal vein 304 delivers about 75% of the liver's blood supply and carries venous blood draining from the spleen, digestive tract, and its associated organs. The hepatic vein 306 returns blood to the heart. In each voxel, there are cells. Some cells 311 may process or take up the contrast agent being used in the imaging. These cells 311 may move the contrast agent into the bile duct 310. The voxel may contain other cells 313 that do not take up the contrast agent effectively. There is also an extracellular space 301 representing fluid and binding molecules that hold all these other components in place. Of course, in some parts of the organ, there may be voxels that are entirely within one type of component, e.g., arteries, veins, or bile ducts. Other voxels have other fractions of voxel elements.

Molecules can move from the blood into the extracellular volume and from there into cells or bile ducts and vice versa. Different molecules diffuse or are transported at different rates depending on their properties and those of the structures or cells involved. PK/PD modeling (pharmacokinetic/pharmacodynamic modeling) is a technique that combines the two classical pharmacological disciplines of pharmacokinetics and pharmacodynamics. It integrates pharmacokinetic and pharmacodynamic model components into one set of mathematical expressions that allow the description of the time course of effect intensity in response to the administration of a certain dosage. In a simple PK/PD model, K1A and K2A are the transport constants out of and into the hepatic artery 302 and associated capillaries, respectively. K1V and K2V are the constants of the portal vein 304. K3 and K4 are the constants of the cells taking up the contrast agent, and K5 and K6 are the transport constants into and back from the bile ducts. The PK/PD model also determines as parameters one or more of the fractional volumes of each voxel occupied by various compartments: arterial blood, portal blood, venous blood, extracellular fluid, extracellular matrix, and various types of cells, e.g., hepatocytes and non-hepatocytes.

In typical MR imaging, the images acquired are of high enough quality that a radiologist can comfortably view them and reliably interpret them to arrive at his or her diagnosis. This typically means that scan times of tens of seconds or even minutes are required to produce a single image.

Technologies have been developed to speed up the creation of MRI images that meet display criteria. These techniques include, for example, parallel imaging, compressed sensing, sparse imaging, and many other techniques. Images created using these techniques may be used in various non-limiting embodiments or aspects of the present disclosure, for example, to determine one or more properties of voxels of a patient's tissue as described herein.

7, in some non-limiting embodiments or aspects, in step 704, a faster but noisier image can be created using an acquisition time of just a few seconds or tens of seconds. Thus, the data captured for each voxel (e.g., measurement information, parameters of a voxel of tissue measured at two or more time points, etc.) can better measure or approximate the shape of curve S in FIG. 1, even if there is some noise in the measurements that may still be annoying or confusing to a human visual observer.

In some non-limiting embodiments or aspects, in step 706 of FIG. 7, the time sequence of the noisier images (e.g., measurement information, parameters of tissue voxels measured at two or more time points, etc.) may be fed to a predictive model (e.g., an artificial neural network, a predictive model as described herein, etc.) that has been trained by supervised learning to predict one or more characteristics (and/or one or more images including or indicative of one or more characteristics) at a time point and/or time period corresponding to one of the two or more time points and/or at a time point and/or time period following the two or more time points (e.g., following the two or more time points) based on the measurement information including parameters of tissue voxels measured at the two or more time points.

In some non-limiting embodiments or aspects, in step 706 of FIG. 7, the time sequence of the noisier images may be used to fit various model parameters to the data (e.g., parameters of tissue voxels measured at two or more time points, etc.). In some non-limiting embodiments or aspects, the PK/PD model of FIG. 8 may be fitted to parameters of tissue voxels measured at two or more time points (e.g., curves measured per voxel, etc.). For example, the image analysis and calculation system 614 may fit a PK/PD model of tissue voxels to parameters of tissue voxels measured at two or more time points and determine one or more properties of the tissue voxels based on the PK/PD model fitted to parameters of tissue voxels measured at two or more time points. As an example, the parameters or required input functions may be measured using voxels that are entirely hepatic arteries and portal veins. There are many curve fitting methods known in the literature. Least squares curve fitting is a simple but not necessarily the most efficient or effective method. What may be advantageous for use in non-limiting embodiments or aspects of the present disclosure is, for simple extracellular imaging (meaning K3-K6 are zero) as a way to measure liver perfusion, using comparison, matching, or fingerprinting to find the best fit in a set of ideal PK/PD curves that are pre-calculated for each of a discrete set of parameters and placed in a database or dictionary, as discussed by Dr. Nicole Seiberlich during RSNA 2018, course RC629C. PK/PD parameters per voxel are found by dictionary lookup, images at the corresponding time and any future time points can be inferred by developing PK/PD equations per voxel and converting concentrations to signal intensities, and PK/PD measures over time, such as drug uptake, distribution, and excretion, allowing certain conclusions on cellular function of the kidney, liver, and blood-brain barrier, can be determined from relative or absolute contrast-induced tissue enhancement over time. For example, the image analysis and calculation system 614 may fit a PK/PD curve of multiple pre-calculated PK/PD curves for a parameter to a parameter of a tissue voxel measured at two or more time points, and determine one or more characteristics of the tissue voxel based on the PK/PD curves fitted to the parameter at the two or more time points.

In PK/PD analysis, an input function is commonly used. For example, the input function can be measured in an image, e.g., an image of the aorta adjacent to the liver. A relatively narrow input function is generally considered to be preferred, using a rapid, short drip or bolus of contrast, a few seconds in length, followed by enough saline to move the bolus through the arm and into the central circulation, as described in U.S. Patent Application No. 16/346,219, the entire contents of which are incorporated herein by reference. A disadvantage of this short contrast bolus is that the contrast bolus spreads as it travels through the patient's central circulation, spreading into a bolus of 10-15 seconds in width. A further disadvantage is that the bolus shape is primarily patient-dependent. Alternatively, a longer bolus, optionally >15 seconds, can be used, as described in U.S. Patent No. 9,867,589, the entire contents of which are incorporated herein by reference. A longer bolus causes the input function to be less variably based on the patient, but rather determined by the duration of the infusion. This may be preferable in some types of analysis, modeling, or curve fitting described herein by limiting the range of parameters that may be expected to occur in the model. One drawback of a longer bolus may be that normal arterial, portal, or other phase images may not be available due to time overlap. If it is desired to present a "standard" image to the radiologist, the image analysis and calculation system 614 may analyze one or more characteristics of the model for each voxel, and a "standard image" may be constructed from the model parameters for that voxel. This longer bolus approach may be particularly beneficial in PET imaging to avoid detector saturation or pulse pile-up and dead space correction effects. It may also be beneficial for CT by allowing fewer scans to be taken since contrast levels are changing at a slower rate, timing may be better predicted, and it is less variable depending on the patient's physiology. Simplified PK/PD models such as Patlak analysis and/or other such models are used when appropriate to the diagnostic question being asked and the characteristics being assessed.

In another non-limiting embodiment or aspect, the curve S is approximated by or decomposed into a set of selected basis functions (e.g., a set of polynomial functions, a set of Laplace functions, a set of Fourier functions, etc.). For example, a basis function is an element of a certain basis in a function space. Just as every vector in a vector space can be expressed as a linear combination of basis vectors, every continuous function in a function space can be expressed as a linear combination of basis functions. As an example, the image analysis and calculation system 614 may use a set of basis functions to approximate a curve representing one or more properties of a tissue voxel, fit the approximated curve to parameters of the tissue voxel measured at two or more time points, and determine one or more properties of the tissue voxel based on the approximated curve fitted to parameters of the tissue voxel measured at two or more time points. See, for example, H. K. et al., "Computational Modeling of 3D Images of a Voxel ... As explained in the article by Thompson et al., “Indicator Transit Time Considered as a Gamma Variate”, Circulation Research, Volume XIV, June 1964, pp502-515, the first-pass contrast enhancement over time curve can be modeled or approximated as a gamma variate curve according to the following equation (1):
C(t)=K(t-AT) ^a . ^* exp(-(t-AT)/b (1)
where t=time after injection, C(t)=concentration at time, K=constant scale factor, AT=emergence time, and a, b=optional parameters for t>AT.

The curve S may be approximated by two or more gamma variate curves to represent the first-pass portion of the A curve, the first-pass portion of the V curve, optionally with a third gamma variate curve to represent steady-state recirculation and redistribution, and one linearly increasing curve C(t)=m(t-AT')+n, where m and n are optional parameters, AT' is the start of the linear rise, and t is the time after injection for t>AT'. However, non-limiting embodiments or aspects are not limited to using gamma variate curves for modeling, and any other set of basis functions (e.g., a set of polynomial functions, a set of Laplace functions, a set of Fourier functions, etc.) may be used for modeling, and any computationally efficient curve fitting program may be used to determine the best-fit parameters for the measured curve S. Once the basis function parameters for each voxel are found through the curve fitting process, the image at the corresponding time and at any time point can be inferred by rolling or rolling back the equation for each voxel to convert the concentration to signal intensity for the desired time t. For example, the image analysis and calculation system 614 can fit one of a plurality of curves pre-calculated for one or more parameters using a set of basis functions to the parameters of the tissue voxel measured at two or more time points, and determine one or more properties of the tissue voxel based on the curve fitted to the parameters of the tissue voxel measured at two or more time points.

As explained elsewhere herein, a longer contrast bolus or injection may cause the curve S to have a slower rise and fall, which may affect or constrain the basis function parameters expected in the model, thereby simplifying, speeding up, or increasing the accuracy of the curve fitting activity and/or the resulting model.

When a model or algorithm, e.g., PK/PD, basis function, AI or other models known to those skilled in the art, is fitted to two or more time points per voxel, images can be created or information derived that is never measured independently in a physical context. The "white liver" in FIG. 9 is a non-limiting example of this. Another non-limiting embodiment or aspect is an image or sequence of images created to show contrast agent flowing through the liver over time as if it entered the liver only through the portal vein and not through the portal artery, or vice versa. Similarly, images can be created that result from the use of a very fast or short bolus, even if a longer bolus was actually used throughout the imaging protocol. Such an ability to create images that were physically impossible to measure given the actual imaging protocol used could lead to new understanding in diagnosis, nuances or health conditions, or treatment and response to disease treatment. This is in addition to currently performed images of the PK/PD parameters themselves, e.g., blood volume passage or various K or complex constant maps.

Still referring to FIG. 7, in some non-limiting embodiments or aspects, in step 708, the image analysis and calculation system 614 may analyze one or more characteristics associated with the voxels of the patient's tissue and/or one or more characteristics associated with the one or more images of the patient's tissue to calculate an absolute value for the washout of the lesion over time compared to the arterial phase enhancement (e.g., liver). Currently, the washout of the lesion is typically assessed visually by comparing the enhancement of the surrounding tissue to the enhancement of the lesion during the contrast phase that follows the arterial phase. However, this exposes the risk of misinterpretation since some contrast agents (liver-specific agent Primovist®) may be taken up by liver cells that surround or are part of the lesion (e.g., potentially a metastasis or other tumor). Such early uptake of Primovist® may create the illusion of a stronger contrast between the lesion and the surrounding tissue, which may be misinterpreted as washout when in fact none is present. This highlights the risk of misdiagnosis without analysis of non-limiting embodiments or aspects of the present disclosure that allows for separate or distinct decomposition and/or visualization of these aspects.

In some non-limiting embodiments or aspects, the examination region includes the liver or a portion of the liver of a mammal (preferably a human). In some non-limiting embodiments or aspects, the examination region may include the lungs or a portion thereof. The lungs receive circulation of deoxygenated blood from the right heart and oxygenated blood from the left heart. In addition, the lungs receive gases through the airways. Thus, the lungs may receive contrast agents, either intravenously or gaseous contrast agents through the airways. Thus, non-limiting embodiments or aspects of the present disclosure may apply to inhaled contrast agents, as well as IV injected contrast agents. In addition, most tissues act as lymphatic systems (e.g., the glymphatic system in the brain) for circulation of fluid through the extracellular space of the tissue. Non-limiting embodiments or aspects of the present disclosure may include the flow of contrast agents, or lack thereof, through this lymphatic and glymphatic system.

In various non-limiting embodiments or aspects described herein, various tissue characteristics and voxel parameters are listed for purposes of understanding and disclosure. It should be understood that these are exemplary and not a limiting or exhaustive list. Other characteristics and/or parameters known in the medical arts may be used. Additionally, characteristics and/or parameters under study or discovered in the future may also benefit from application of non-limiting embodiments or aspects of the present disclosure in processing, imaging, and/or analysis of data from an imaging study and/or series of studies.

Claims (8)

obtaining measurement information associated with parameters of voxels of an image of tissue of a patient , the measurement information being measured at two or more time points to provide first measurement information associated with the parameters of the voxels at a first one of the two or more time points and second measurement information associated with the parameters of the voxels at a second one of the two or more time points, the two or more time points occurring before one or more tissue properties of the voxels are separable or identifiable in another image generated based on the parameters of the voxels measured at one time point;
The one or more tissue characteristics include:
Concentration of contrast agent in the arteries,
Concentration of intravenous contrast agent,
the concentration of contrast agent in the cells,
the sum of the arterial, venous, and intracellular concentrations of contrast agent;
one or more pharmacokinetic model parameters associated with the movement of the contrast agent through the tissue space; or
Any combination of these may be included.
determining the one or more characteristics of the voxel of the image of the tissue based on the first measurement information associated with the parameter of the voxel at the first time point, the second measurement information associated with the parameter of the voxel at the second time point, and a desired rate of change and/or plateau level of concentration of contrast agent delivered to the patient .
Computer-implemented method. The computer-implemented method of claim 1 , wherein the one or more properties of the voxels of the image of the tissue are determined for a time point subsequent to the two or more time points. 3. The computer-implemented method of claim 2, further comprising generating one or more images comprising one or more properties of the voxels of the image of the tissue at the time points after the two or more time points based on the one or more tissue properties. Determining the one or more properties of the voxels of the image of the tissue comprises:
3. The computer-implemented method of claim 2, further comprising: providing the measurement information associated with the parameters of the voxels of the image of the tissue of the patient to a predictive model, the predictive model being trained by supervised learning to predict the one or more properties of the voxels of the image of the tissue based on the measurement information associated with the parameters at the two or more time points. 1. A computer program product comprising at least one non-transitory computer-readable medium containing program instructions ,
The program instructions, when executed by at least one processor, cause the at least one processor to:
acquiring measurement information associated with parameters of voxels of an image of tissue of a patient , the measurement information being measured at two or more time points to provide first measurement information associated with the parameters of the voxels at a first one of the two or more time points and second measurement information associated with the parameters of the voxels at a second one of the two or more time points, the two or more time points occurring before one or more tissue properties of the voxels are separable or identifiable in another image generated based on the parameters of the voxels measured at one time point;
The one or more tissue characteristics include:
Concentration of contrast agent in the arteries,
Concentration of intravenous contrast agent,
the concentration of contrast agent in the cells,
the sum of the arterial, venous, and intracellular concentrations of contrast agent;
one or more pharmacokinetic model parameters associated with the movement of the contrast agent through the tissue space; or
Any combination of these may be included.
The program instructions, when executed by the at least one processor, cause the at least one processor to:
determining the one or more characteristics of the voxel of the image of the tissue based on the first measurement information associated with the parameter of the voxel at the first time point, the second measurement information associated with the parameter of the voxel at the second time point, and a desired rate of change and/or plateau level of concentration of contrast agent delivered to the patient .
Computer program products. The computer program product of claim 5 , wherein the one or more properties of the voxels of the image of the tissue are determined for a time point subsequent to the two or more time points. The instructions cause the at least one processor to:
7. The computer program product of claim 6, further comprising: generating one or more images comprising one or more properties of the voxels of the image of the tissue at the time point after the two or more time points based on the one or more tissue properties. The instructions cause the at least one processor to:
7. The computer program product of claim 6, further comprising: determining the one or more properties of the voxels of the image of the tissue of the patient by providing the measurement information associated with the parameters of the voxels of the image of the tissue to a predictive model, the predictive model being trained by supervised learning to predict the one or more properties of the voxels of the image of the tissue based on the measurement information associated with the parameters at the two or more time points.

Images